Design And Implementation Of A Scheme About Fully Homomorphic Encryption

The fully homomorphic encryption algorithm can achieve arbitrary complex computation for data in ciphertext form,which can guarantee the confidentiality of user data in cloud computing scenarios.The current research in the field of fully homomorphic encryption can be roughly divided into three levels.The research and optimization of the underlying mathematical operations required by the fully homomorphic encryption algorithm is the first layer,the research and improvement of the fully homomorphic encryption algorithm is the second layer,and the research and improvement of the fully homomorphic encryption transpiler is the third layer.The scheme about fully homomorphic encryption proposed in this paper is in the third layer,which is about how programmer users can better use the fully homomorphic encryption transpiler.The main achievements and innovations of this paper are as follows:(1)Currently existing fully homomorphic encryption transpilers generally do not support transpile recursion functions.To address this problem,this paper proposes a scheme for transpile recursive functions into loops which can be used with fully homomorphic encryption transpilers.The scheme,for a fully homomorphic encryption transpiler capable of transpile non-functional programming languages,improves on the state machine based recursive to loop framework,making the functions before and after the conversion more similar in branching structure,and making the conversion process of converting recursive functions to a loop implementation that simulates a system stack easier to perform.For functional programming languages,the scheme adds support for define expressions in the implementation of automatic CPS transform functions,thus enabling automatic CPS transformations of recursive functions,and the output of the CPS transformations can eventually be transformed into a loop implementation by the loop computation framework in the scheme.(2)In order to address the problem that the implementation information of user-defined algorithms may be leaked due to the use of fully homomorphic encryption transpilers in cloud computing,this paper first discusses the possibility of hiding the implementation information of algorithms from a theoretical point of view,and then proposes a scheme to hide the implementation information of user-defined algorithms from third-party cloud servers based on this theory.This scheme can guarantee the confidentiality of the information about the implementation of userdefined algorithms and the confidentiality of the input data for the user’s algorithms.This scheme uses a fully homomorphic encryption transpiler to implement a homomorphic interpreter for a programming language from the top down,which is significantly less complex than implementing a homomorphic interpreter from the bottom up,and makes it easier to modify and maintain the homomorphic interpreter later.After that,this paper verifies the practicability of the scheme through experiments,and obtains the effect of introducing an interpreter into the fully homomorphic encryption transpiler on the running efficiency of the user program through comparison experiments.Finally,this paper discusses the application of the scheme in verifying the homomorphic computation results returned by the cloud server.